A portion of the disclosure of this patent document contains material, which is subject to mask work protection. The mask work owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all mask work rights whatsoever.
This invention relates to semiconductor devices, methods of manufacturing semiconductor devices, and more particularly, to Zener diodes and methods of manufacturing high-yield, high-precision Zener diodes.
Semiconductor devices are critical components in a vast array of modern products. Purified crystalline semiconductor materials have highly useful electrical characteristics when specific types of impurities, called dopants, are introduced into the semiconductor""s crystalline structure. Depending on the nature of the dopant introduced into the semiconductor material, the semiconductor material will take on a particular conductivity type, either P-type or N-type. In connection with the present invention, to say that a material has one conductivity type that is consistent with another conductivity type means that both conductivity types are either P-type or N-type.
The element silicon is a an example of a semiconductor material. Pure silicon has four electrons in the valence shell of each of its atoms. Pure crystalline silicon forms a lattice structure, in which silicon""s valence electrons form stable covalent bonds with other silicon atoms.
An example of P-type material is pure silicon doped with an impurity such as boron, aluminum, gallium, or indium. These materials are referred to as xe2x80x9cacceptorxe2x80x9d impurities because their valence shells contain only three electrons. When these materials are introduced into a semiconductor crystal, the uniform lattice structure of the silicon is affected because the three-electron valence shells of the doping material can""t complete the lattice. The vacancy created by the lack of a fourth electron is called a hole. Holes are loosely held to the impurity atom so that, when affected by an electric field, electrons can drift into the hole, thus causing the hole to appear to drift. In this way, the hole acts as a positive-charge current carrier.
An example of N-type material is pure silicon doped with a very small amount of impurities containing five electrons in the valence shell. These materials can be antimony, phosphorus, or arsenic. Because of their extra electrons, they are called donor impurities. When these materials are blended with pure silicon, the uniform lattice structure of the silicon is affected because the five-electron valence shell of the doping material has too many electrons to simply complete the lattice structure. These extra electrons are loosely held to their impurity atoms so that, when affected by an electric field, the electrons can drift, thus acting as a negative-charge current carrier.
When a junction is formed between P material and N material, (xe2x80x9cP/N junctionxe2x80x9d) the extra current carriers tend to cross the junction so that the lattice structure in the vicinity of the junction tends to have four electrons associated with each atom. The region where this phenomenon occurs is called the depletion region since both the P- and N-type materials have been depleted of their current carriers in this region.
In a P/N junction device, sometimes called a diode or rectifier, the electrode connected to the P-type material is referred to as the anode, and the electrode connected to the N-type material is called the cathode. The depletion region of a P/N junction has the useful property of causing a P/N junction device to conduct current when a positive voltage (above a forward voltage drop threshold) is applied across the P/N junction and to block the flow of current when a negative voltage is applied across the P/N junction. A positive voltage applied from anode to cathode is referred to as forward bias, and a negative voltage is referred to as reverse bias.
Accordingly, diodes conduct current from anode to cathode at forward biased voltages above the forward voltage drop, and diodes block current under reverse bias up to a point at which the diodes break down under a sufficiently high reverse biased voltage. Diodes that take advantage of breakdown characteristics are called Zener diodes.
Zener diodes have been used since the late 1950""s as voltage references or for voltage regulation, originally as an alternative to the vacuum tube. Zener diodes have the useful property of blocking current under reverse bias, up to a threshold or breakdown voltage. When installed in parallel with a load, reverse biased Zener diodes clamp the voltage across the load at the Zener diode""s breakdown voltage.
Zener diodes are P/N junction devices that are designed to operate nondestructively in reverse bias breakdown mode. While every P/N junction will break down under a sufficiently high reverse bias, a low-power rectifying diode will break down at a fairly high voltage and would likely be damaged by the resulting current. However, Zener diodes are designed to operate in breakdown mode, at specified currents, without sustaining damage.
A relatively lightly doped P/N junction will exhibit avalanche breakdown at the relatively high voltage of approximately 30V to 50V. Avalanche breakdown is the result of energizing thermally produced electron/hole pairs in the depletion region surrounding a P/N junction with the electric field associated with the reverse biased P/N junction. Given a sufficiently large electric field, energized electrons eventually take on enough energy to ionize atoms of the semiconductor material in the depletion region. Next, electrons that are released by ionization themselves become energized by the electric field, resulting in further ionization. The result of the chain reaction of ionization is the occurrence of sufficient numbers of charge carriers to enable the P/N junction to conduct electrical current. Observers have remarked that this chain reaction is like an avalanche on a snow covered mountain. Accordingly this type of breakdown is called avalanche breakdown.
Zener diodes are designed to break down at a specific voltage with a sharp reproducible characteristic. The diodes are designed to conduct the breakdown current nondestructively.
Zener diodes are generally used either as a voltage reference or as transient voltage suppressors. When used as voltage references, a high degree of precision is required for some electrical circuit designs. Accordingly, Zener diodes are frequently specified in terms of a xc2x1 percentage error in breakdown voltage tolerance.
Known processes of manufacturing Zener diodes consist of fabricating them on thinly sliced wafers of crystalline semiconductor substrate. Conventional substrate wafers are formed with a high purity, monocrystalline, semiconductor material by a known monocrystalline growth method. In the growth method, a pool of doped molten liquid semiconductor material is seeded with a small semiconductor crystal. The seed is slowly drawn out of the pool, and as it is drawn out, the molten semiconductor atoms or molecules align with the lattice structure of the seed crystal to form a generally cylindrical ingot of semiconductor material. The crystalline semiconductor material can also be fabricated with known float zone methods. The ingot is then sliced into generally circular substrate wafers, of a conductivity type determined by the dopant type and concentration introduced into the molten semiconductor.
Ideally, each such semiconductor wafer has precisely the same doping concentration and resistivity. However, in practice, this is not the case. Because of inherent properties of dopant materials and the way the dopant materials are introduced into the semiconductor material there are differences in dopant concentration and resistivity along the length of a semiconductor ingot. Further, there are also differences in dopant concentration and resistivity across each wafer.
There are several known processes for modifying the physical properties and conductivity properties of various regions of a substrate wafer to fabricate a semiconductor device. Diffusion is the process of heating the substrate in the presence of a material containing atoms to be diffused into the substrate. For instance, a conventional way to fabricate a P-type layer on an N-type substrate is to use planar dopant sources. Silicon wafers are heated facing the dopant source wafers. Over time, a layer of B2O3 covers the silicon wafers and boron diffuses into the N-type substrate creating a layer of P-type material. Because of the high temperatures at which the diffusion process is performed, the boron atoms are able to replace silicon atoms in the silicon crystal structure. Parameters including the concentration of the dopant gas and the amount of time of the diffusion control the depth of the layer and the concentration of the dopant.
Another way to form a layer on a substrate is using epitaxy. Epitaxial deposition methods involve growing a layer of material by gradually adding a combination of silicon and dopant atoms to the surface of the substrate so that the added atoms maintain the same crystal structure as the substrate.
Another way to form a layer in an existing semiconductor material is by way of ion implantation. In ion implantation, individual ions of dopant materials are accelerated to great speeds and shot into the semiconductor material. By altering the energy at which ions are implanted, the depth of implantation can be controlled. Further, through ion implantation, the concentration or dose of dopants to be introduced can be closely controlled. As contrasted with diffusion, ion implantation can be performed without heating the substrate to high temperatures.
In conventional Zener diode fabrication processes, the breakdown voltage is determined both by the resistivity of the substrate and the resistivity of a semiconductor layer fabricated on the surface of the substrate, using for example diffusion. Since the resistivity of a substrate wafer varies from wafer to wafer and from point to point on a particular wafer, the breakdown voltage of conventionally fabricated Zener diodes also varies. These variations cannot be compensated for, for example, by differences in diffusion time, because of the initial non-uniformity of the wafers.
As a result, Zener diodes fabricated on a single wafer using known processes have different threshold voltages depending on the initial resistivity of the corresponding portion of the wafer. Even at larger target breakdown voltage tolerances, e.g. xc2x15%, the yield loss caused by missing the breakdown voltage target is significant, typically around 20%.
In addition to differences in substrate resistivity, another problem with conventional Zener diode fabrcation involves differences in breakdown voltages because of variations in the behavior of P/N junction breakdown at the physical edge of a semiconductor device. Since avalanche breakdown is closely related to the presence and orientation of an electrical field at the P/N junction, unpredictable variations in Zener diode breakdown voltage result from differences in the electrical field at the physical edges of the semiconductor device. Accordingly, separation of the active portion of the device from the edge areas, or termination of the active area of the Zener diode, is useful to create a region in which a predictable electric field will be present at a particular reverse bias.
Known systems either provide complex semiconductor structures to terminate the region in which the breakdown occurs or provide for no termination at all and test for acceptable devices from a pool of devices resulting in a low-yield fabrication process.
There is, therefore, provided in the practice of the invention a novel semiconductor device, which can be efficiently fabricated for use in Zener diode applications. Precision Zener diodes, methods for manufacturing precision Zener diodes, and consumer electronics employing the Zener diodes are provided. The Zener diodes are made from a semiconductor substrate layer on which is grown an epitaxial layer. The epitaxial layer has a resistivity greater than that of the substrate. The diode also has an interior region of doped semiconductor material of the same conductivity type as the substrate. The interior region extends through the epitaxial layer and into the substrate layer. The diode also has a junction layer of a conductivity type different from the substrate. The junction layer is formed in the epitaxial surface, and the junction layer forms an interior P/N junction with the interior region and a peripheral P/N junction with a peripheral portion of the device.
In an exemplary embodiment, an N-type substrate is used. The substrate layer can be doped with arsenic or antimony. The epitaxial layer, deposited on the substrate, can be of a thickness ranging from about 6 microns to about 15 microns. The substrate resistivity can be in the range of about 1E-3 ohm-cm to about 5E-3 ohm-cm.
In one embodiment, the interior region is produced by ion implantation of phosphorous, and the junction layer is produced by ion implantation of boron. An additional process can optionally be performed, in which a low contact resistance layer is implanted into an exterior surface of the junction layer.
In one embodiment, the interior P/N junction has an interior breakdown voltage lower than an peripheral breakdown voltage associated with the peripheral P/N junction.